Integer linear optimization: Cutting Planes 2: Unterschied zwischen den Versionen

Aus Operations-Research-Wiki
Wechseln zu: Navigation, Suche
[unmarkierte Version][unmarkierte Version]
(Approach)
(Approach)
Zeile 17: Zeile 17:
 
:Row r:
 
:Row r:
 
::<math>BV_r+\sum (a_{ij}*NBV_j) = b_r</math>
 
::<math>BV_r+\sum (a_{ij}*NBV_j) = b_r</math>
:Converting r to: You have to break all <math>a_{rj}</math> and <math>b_j</math> into the integer part and the rest.
+
:Converting r to: You have to break all <math>a_{rj}</math> and <math>b_j</math> into the integer part <math>g_{rj}</math> and the rest <math>f_j</math>.

Version vom 20. Juni 2013, 12:07 Uhr

Cutting Planes Gruppe 2

Idea

The Idea of the Cutting Plane is to add Restrictions (the Cutting Planes) to contract the solution space more and more to get integers. These restrictions cut of the non-integer parts of the solution.


Approach

First of all you should divide your restrictions trough the greates common factor.

Example: 6x1 – 15x2 <= 120
Fehler beim Parsen (http://mathoid.testme.wmflabs.org Serverantwort ist ungültiges JSON.): \Rightarrow
3 x1 – 5 x2 <= 40
(1) You search for the continous Optimum with the Simplex Algorithm. If your solution is integer you are finished.
(2) Otherwise you have to insert Cutting Planes
Your source row is the one with the greates non integer part.
Row r:
Fehler beim Parsen (http://mathoid.testme.wmflabs.org Serverantwort ist ungültiges JSON.): BV_r+\sum (a_{ij}*NBV_j) = b_r
Converting r to: You have to break all and into the integer part and the rest .